summaryrefslogtreecommitdiff
path: root/fs/erofs/zdata.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/erofs/zdata.c')
-rw-r--r--fs/erofs/zdata.c456
1 files changed, 212 insertions, 244 deletions
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index f416b73e0ca3..e3f28a1bb945 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -12,12 +12,6 @@
#define Z_EROFS_PCLUSTER_MAX_PAGES (Z_EROFS_PCLUSTER_MAX_SIZE / PAGE_SIZE)
#define Z_EROFS_INLINE_BVECS 2
-/*
- * let's leave a type here in case of introducing
- * another tagged pointer later.
- */
-typedef void *z_erofs_next_pcluster_t;
-
struct z_erofs_bvec {
struct page *page;
int offset;
@@ -48,10 +42,10 @@ struct z_erofs_pcluster {
struct lockref lockref;
/* A: point to next chained pcluster or TAILs */
- z_erofs_next_pcluster_t next;
+ struct z_erofs_pcluster *next;
- /* I: start block address of this pcluster */
- erofs_off_t index;
+ /* I: start physical position of this pcluster */
+ erofs_off_t pos;
/* L: the maximum decompression size of this round */
unsigned int length;
@@ -79,12 +73,12 @@ struct z_erofs_pcluster {
/* I: compression algorithm format */
unsigned char algorithmformat;
+ /* I: whether compressed data is in-lined or not */
+ bool from_meta;
+
/* L: whether partial decompression or not */
bool partial;
- /* L: indicate several pageofs_outs or not */
- bool multibases;
-
/* L: whether extra buffer allocations are best-effort */
bool besteffort;
@@ -94,12 +88,11 @@ struct z_erofs_pcluster {
/* the end of a chain of pclusters */
#define Z_EROFS_PCLUSTER_TAIL ((void *) 0x700 + POISON_POINTER_DELTA)
-#define Z_EROFS_PCLUSTER_NIL (NULL)
struct z_erofs_decompressqueue {
struct super_block *sb;
+ struct z_erofs_pcluster *head;
atomic_t pending_bios;
- z_erofs_next_pcluster_t head;
union {
struct completion done;
@@ -109,14 +102,9 @@ struct z_erofs_decompressqueue {
bool eio, sync;
};
-static inline bool z_erofs_is_inline_pcluster(struct z_erofs_pcluster *pcl)
-{
- return !pcl->index;
-}
-
static inline unsigned int z_erofs_pclusterpages(struct z_erofs_pcluster *pcl)
{
- return PAGE_ALIGN(pcl->pclustersize) >> PAGE_SHIFT;
+ return PAGE_ALIGN(pcl->pageofs_in + pcl->pclustersize) >> PAGE_SHIFT;
}
static bool erofs_folio_is_managed(struct erofs_sb_info *sbi, struct folio *fo)
@@ -140,7 +128,7 @@ struct z_erofs_pcluster_slab {
static struct z_erofs_pcluster_slab pcluster_pool[] __read_mostly = {
_PCLP(1), _PCLP(4), _PCLP(16), _PCLP(64), _PCLP(128),
- _PCLP(Z_EROFS_PCLUSTER_MAX_PAGES)
+ _PCLP(Z_EROFS_PCLUSTER_MAX_PAGES + 1)
};
struct z_erofs_bvec_iter {
@@ -274,7 +262,6 @@ static struct z_erofs_pcluster *z_erofs_alloc_pcluster(unsigned int size)
pcl = kmem_cache_zalloc(pcs->slab, GFP_KERNEL);
if (!pcl)
return ERR_PTR(-ENOMEM);
- pcl->pclustersize = size;
return pcl;
}
return ERR_PTR(-EINVAL);
@@ -301,6 +288,7 @@ static struct workqueue_struct *z_erofs_workqueue __read_mostly;
#ifdef CONFIG_EROFS_FS_PCPU_KTHREAD
static struct kthread_worker __rcu **z_erofs_pcpu_workers;
+static atomic_t erofs_percpu_workers_initialized = ATOMIC_INIT(0);
static void erofs_destroy_percpu_workers(void)
{
@@ -320,7 +308,7 @@ static void erofs_destroy_percpu_workers(void)
static struct kthread_worker *erofs_init_percpu_worker(int cpu)
{
struct kthread_worker *worker =
- kthread_create_worker_on_cpu(cpu, 0, "erofs_worker/%u", cpu);
+ kthread_run_worker_on_cpu(cpu, 0, "erofs_worker/%u");
if (IS_ERR(worker))
return worker;
@@ -346,12 +334,8 @@ static int erofs_init_percpu_workers(void)
}
return 0;
}
-#else
-static inline void erofs_destroy_percpu_workers(void) {}
-static inline int erofs_init_percpu_workers(void) { return 0; }
-#endif
-#if defined(CONFIG_HOTPLUG_CPU) && defined(CONFIG_EROFS_FS_PCPU_KTHREAD)
+#ifdef CONFIG_HOTPLUG_CPU
static DEFINE_SPINLOCK(z_erofs_pcpu_worker_lock);
static enum cpuhp_state erofs_cpuhp_state;
@@ -408,17 +392,56 @@ static void erofs_cpu_hotplug_destroy(void)
if (erofs_cpuhp_state)
cpuhp_remove_state_nocalls(erofs_cpuhp_state);
}
-#else /* !CONFIG_HOTPLUG_CPU || !CONFIG_EROFS_FS_PCPU_KTHREAD */
+#else /* !CONFIG_HOTPLUG_CPU */
static inline int erofs_cpu_hotplug_init(void) { return 0; }
static inline void erofs_cpu_hotplug_destroy(void) {}
-#endif
+#endif/* CONFIG_HOTPLUG_CPU */
+static int z_erofs_init_pcpu_workers(struct super_block *sb)
+{
+ int err;
-void z_erofs_exit_subsystem(void)
+ if (atomic_xchg(&erofs_percpu_workers_initialized, 1))
+ return 0;
+
+ err = erofs_init_percpu_workers();
+ if (err) {
+ erofs_err(sb, "per-cpu workers: failed to allocate.");
+ goto err_init_percpu_workers;
+ }
+
+ err = erofs_cpu_hotplug_init();
+ if (err < 0) {
+ erofs_err(sb, "per-cpu workers: failed CPU hotplug init.");
+ goto err_cpuhp_init;
+ }
+ erofs_info(sb, "initialized per-cpu workers successfully.");
+ return err;
+
+err_cpuhp_init:
+ erofs_destroy_percpu_workers();
+err_init_percpu_workers:
+ atomic_set(&erofs_percpu_workers_initialized, 0);
+ return err;
+}
+
+static void z_erofs_destroy_pcpu_workers(void)
{
+ if (!atomic_xchg(&erofs_percpu_workers_initialized, 0))
+ return;
erofs_cpu_hotplug_destroy();
erofs_destroy_percpu_workers();
+}
+#else /* !CONFIG_EROFS_FS_PCPU_KTHREAD */
+static inline int z_erofs_init_pcpu_workers(struct super_block *sb) { return 0; }
+static inline void z_erofs_destroy_pcpu_workers(void) {}
+#endif/* CONFIG_EROFS_FS_PCPU_KTHREAD */
+
+void z_erofs_exit_subsystem(void)
+{
+ z_erofs_destroy_pcpu_workers();
destroy_workqueue(z_erofs_workqueue);
z_erofs_destroy_pcluster_pool();
+ z_erofs_crypto_disable_all_engines();
z_erofs_exit_decompressor();
}
@@ -440,19 +463,8 @@ int __init z_erofs_init_subsystem(void)
goto err_workqueue_init;
}
- err = erofs_init_percpu_workers();
- if (err)
- goto err_pcpu_worker;
-
- err = erofs_cpu_hotplug_init();
- if (err < 0)
- goto err_cpuhp_init;
return err;
-err_cpuhp_init:
- erofs_destroy_percpu_workers();
-err_pcpu_worker:
- destroy_workqueue(z_erofs_workqueue);
err_workqueue_init:
z_erofs_destroy_pcluster_pool();
err_pcluster_pool:
@@ -462,39 +474,32 @@ err_decompressor:
}
enum z_erofs_pclustermode {
+ /* It has previously been linked into another processing chain */
Z_EROFS_PCLUSTER_INFLIGHT,
/*
- * a weak form of Z_EROFS_PCLUSTER_FOLLOWED, the difference is that it
- * could be dispatched into bypass queue later due to uptodated managed
- * pages. All related online pages cannot be reused for inplace I/O (or
- * bvpage) since it can be directly decoded without I/O submission.
+ * A weaker form of Z_EROFS_PCLUSTER_FOLLOWED; the difference is that it
+ * may be dispatched to the bypass queue later due to uptodated managed
+ * folios. All file-backed folios related to this pcluster cannot be
+ * reused for in-place I/O (or bvpage) since the pcluster may be decoded
+ * in a separate queue (and thus out of order).
*/
Z_EROFS_PCLUSTER_FOLLOWED_NOINPLACE,
/*
- * The pcluster was just linked to a decompression chain by us. It can
- * also be linked with the remaining pclusters, which means if the
- * processing page is the tail page of a pcluster, this pcluster can
- * safely use the whole page (since the previous pcluster is within the
- * same chain) for in-place I/O, as illustrated below:
- * ___________________________________________________
- * | tail (partial) page | head (partial) page |
- * | (of the current pcl) | (of the previous pcl) |
- * |___PCLUSTER_FOLLOWED___|_____PCLUSTER_FOLLOWED_____|
- *
- * [ (*) the page above can be used as inplace I/O. ]
+ * The pcluster has just been linked to our processing chain.
+ * File-backed folios (except for the head page) related to it can be
+ * used for in-place I/O (or bvpage).
*/
Z_EROFS_PCLUSTER_FOLLOWED,
};
-struct z_erofs_decompress_frontend {
+struct z_erofs_frontend {
struct inode *const inode;
struct erofs_map_blocks map;
struct z_erofs_bvec_iter biter;
struct page *pagepool;
struct page *candidate_bvpage;
- struct z_erofs_pcluster *pcl;
- z_erofs_next_pcluster_t owned_head;
+ struct z_erofs_pcluster *pcl, *head;
enum z_erofs_pclustermode mode;
erofs_off_t headoffset;
@@ -503,11 +508,11 @@ struct z_erofs_decompress_frontend {
unsigned int icur;
};
-#define DECOMPRESS_FRONTEND_INIT(__i) { \
- .inode = __i, .owned_head = Z_EROFS_PCLUSTER_TAIL, \
- .mode = Z_EROFS_PCLUSTER_FOLLOWED }
+#define Z_EROFS_DEFINE_FRONTEND(fe, i, ho) struct z_erofs_frontend fe = { \
+ .inode = i, .head = Z_EROFS_PCLUSTER_TAIL, \
+ .mode = Z_EROFS_PCLUSTER_FOLLOWED, .headoffset = ho }
-static bool z_erofs_should_alloc_cache(struct z_erofs_decompress_frontend *fe)
+static bool z_erofs_should_alloc_cache(struct z_erofs_frontend *fe)
{
unsigned int cachestrategy = EROFS_I_SB(fe->inode)->opt.cache_strategy;
@@ -524,19 +529,18 @@ static bool z_erofs_should_alloc_cache(struct z_erofs_decompress_frontend *fe)
return false;
}
-static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
+static void z_erofs_bind_cache(struct z_erofs_frontend *fe)
{
struct address_space *mc = MNGD_MAPPING(EROFS_I_SB(fe->inode));
struct z_erofs_pcluster *pcl = fe->pcl;
unsigned int pclusterpages = z_erofs_pclusterpages(pcl);
bool shouldalloc = z_erofs_should_alloc_cache(fe);
- bool standalone = true;
- /*
- * optimistic allocation without direct reclaim since inplace I/O
- * can be used if low memory otherwise.
- */
+ pgoff_t poff = pcl->pos >> PAGE_SHIFT;
+ bool may_bypass = true;
+ /* Optimistic allocation, as in-place I/O can be used as a fallback */
gfp_t gfp = (mapping_gfp_mask(mc) & ~__GFP_DIRECT_RECLAIM) |
__GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN;
+ struct folio *folio, *newfolio;
unsigned int i;
if (i_blocksize(fe->inode) != PAGE_SIZE ||
@@ -544,47 +548,42 @@ static void z_erofs_bind_cache(struct z_erofs_decompress_frontend *fe)
return;
for (i = 0; i < pclusterpages; ++i) {
- struct page *page, *newpage;
-
/* Inaccurate check w/o locking to avoid unneeded lookups */
if (READ_ONCE(pcl->compressed_bvecs[i].page))
continue;
- page = find_get_page(mc, pcl->index + i);
- if (!page) {
- /* I/O is needed, no possible to decompress directly */
- standalone = false;
+ folio = filemap_get_folio(mc, poff + i);
+ if (IS_ERR(folio)) {
+ may_bypass = false;
if (!shouldalloc)
continue;
/*
- * Try cached I/O if allocation succeeds or fallback to
- * in-place I/O instead to avoid any direct reclaim.
+ * Allocate a managed folio for cached I/O, or it may be
+ * then filled with a file-backed folio for in-place I/O
*/
- newpage = erofs_allocpage(&fe->pagepool, gfp);
- if (!newpage)
+ newfolio = filemap_alloc_folio(gfp, 0);
+ if (!newfolio)
continue;
- set_page_private(newpage, Z_EROFS_PREALLOCATED_PAGE);
+ newfolio->private = Z_EROFS_PREALLOCATED_FOLIO;
+ folio = NULL;
}
spin_lock(&pcl->lockref.lock);
if (!pcl->compressed_bvecs[i].page) {
- pcl->compressed_bvecs[i].page = page ? page : newpage;
+ pcl->compressed_bvecs[i].page =
+ folio_page(folio ?: newfolio, 0);
spin_unlock(&pcl->lockref.lock);
continue;
}
spin_unlock(&pcl->lockref.lock);
-
- if (page)
- put_page(page);
- else if (newpage)
- erofs_pagepool_add(&fe->pagepool, newpage);
+ folio_put(folio ?: newfolio);
}
/*
- * don't do inplace I/O if all compressed pages are available in
- * managed cache since it can be moved to the bypass queue instead.
+ * Don't perform in-place I/O if all compressed pages are available in
+ * the managed cache, as the pcluster can be moved to the bypass queue.
*/
- if (standalone)
+ if (may_bypass)
fe->mode = Z_EROFS_PCLUSTER_FOLLOWED_NOINPLACE;
}
@@ -596,7 +595,7 @@ static int erofs_try_to_free_all_cached_folios(struct erofs_sb_info *sbi,
struct folio *folio;
int i;
- DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
+ DBG_BUGON(pcl->from_meta);
/* Each cached folio contains one page unless bs > ps is supported */
for (i = 0; i < pclusterpages; ++i) {
if (pcl->compressed_bvecs[i].page) {
@@ -628,7 +627,7 @@ static bool z_erofs_cache_release_folio(struct folio *folio, gfp_t gfp)
ret = false;
spin_lock(&pcl->lockref.lock);
if (pcl->lockref.count <= 0) {
- DBG_BUGON(z_erofs_is_inline_pcluster(pcl));
+ DBG_BUGON(pcl->from_meta);
for (; bvec < end; ++bvec) {
if (bvec->page && page_folio(bvec->page) == folio) {
bvec->page = NULL;
@@ -665,39 +664,49 @@ static const struct address_space_operations z_erofs_cache_aops = {
.invalidate_folio = z_erofs_cache_invalidate_folio,
};
-int erofs_init_managed_cache(struct super_block *sb)
+int z_erofs_init_super(struct super_block *sb)
{
- struct inode *const inode = new_inode(sb);
+ struct inode *inode;
+ int err;
+ err = z_erofs_init_pcpu_workers(sb);
+ if (err)
+ return err;
+
+ inode = new_inode(sb);
if (!inode)
return -ENOMEM;
-
set_nlink(inode, 1);
inode->i_size = OFFSET_MAX;
inode->i_mapping->a_ops = &z_erofs_cache_aops;
mapping_set_gfp_mask(inode->i_mapping, GFP_KERNEL);
EROFS_SB(sb)->managed_cache = inode;
+ xa_init(&EROFS_SB(sb)->managed_pslots);
return 0;
}
/* callers must be with pcluster lock held */
-static int z_erofs_attach_page(struct z_erofs_decompress_frontend *fe,
+static int z_erofs_attach_page(struct z_erofs_frontend *fe,
struct z_erofs_bvec *bvec, bool exclusive)
{
struct z_erofs_pcluster *pcl = fe->pcl;
int ret;
if (exclusive) {
- /* give priority for inplaceio to use file pages first */
- spin_lock(&pcl->lockref.lock);
- while (fe->icur > 0) {
- if (pcl->compressed_bvecs[--fe->icur].page)
- continue;
- pcl->compressed_bvecs[fe->icur] = *bvec;
+ /* Inplace I/O is limited to one page for uncompressed data */
+ if (pcl->algorithmformat < Z_EROFS_COMPRESSION_MAX ||
+ fe->icur <= 1) {
+ /* Try to prioritize inplace I/O here */
+ spin_lock(&pcl->lockref.lock);
+ while (fe->icur > 0) {
+ if (pcl->compressed_bvecs[--fe->icur].page)
+ continue;
+ pcl->compressed_bvecs[fe->icur] = *bvec;
+ spin_unlock(&pcl->lockref.lock);
+ return 0;
+ }
spin_unlock(&pcl->lockref.lock);
- return 0;
}
- spin_unlock(&pcl->lockref.lock);
/* otherwise, check if it can be used as a bvpage */
if (fe->mode >= Z_EROFS_PCLUSTER_FOLLOWED &&
@@ -727,35 +736,30 @@ static bool z_erofs_get_pcluster(struct z_erofs_pcluster *pcl)
return true;
}
-static int z_erofs_register_pcluster(struct z_erofs_decompress_frontend *fe)
+static int z_erofs_register_pcluster(struct z_erofs_frontend *fe)
{
struct erofs_map_blocks *map = &fe->map;
struct super_block *sb = fe->inode->i_sb;
struct erofs_sb_info *sbi = EROFS_SB(sb);
- bool ztailpacking = map->m_flags & EROFS_MAP_META;
struct z_erofs_pcluster *pcl, *pre;
+ unsigned int pageofs_in;
int err;
- if (!(map->m_flags & EROFS_MAP_ENCODED) ||
- (!ztailpacking && !erofs_blknr(sb, map->m_pa))) {
- DBG_BUGON(1);
- return -EFSCORRUPTED;
- }
-
- /* no available pcluster, let's allocate one */
- pcl = z_erofs_alloc_pcluster(map->m_plen);
+ pageofs_in = erofs_blkoff(sb, map->m_pa);
+ pcl = z_erofs_alloc_pcluster(pageofs_in + map->m_plen);
if (IS_ERR(pcl))
return PTR_ERR(pcl);
- spin_lock_init(&pcl->lockref.lock);
- pcl->lockref.count = 1; /* one ref for this request */
+ lockref_init(&pcl->lockref); /* one ref for this request */
pcl->algorithmformat = map->m_algorithmformat;
+ pcl->pclustersize = map->m_plen;
pcl->length = 0;
pcl->partial = true;
-
- /* new pclusters should be claimed as type 1, primary and followed */
- pcl->next = fe->owned_head;
+ pcl->next = fe->head;
+ pcl->pos = map->m_pa;
+ pcl->pageofs_in = pageofs_in;
pcl->pageofs_out = map->m_la & ~PAGE_MASK;
+ pcl->from_meta = map->m_flags & EROFS_MAP_META;
fe->mode = Z_EROFS_PCLUSTER_FOLLOWED;
/*
@@ -765,13 +769,10 @@ static int z_erofs_register_pcluster(struct z_erofs_decompress_frontend *fe)
mutex_init(&pcl->lock);
DBG_BUGON(!mutex_trylock(&pcl->lock));
- if (ztailpacking) {
- pcl->index = 0; /* which indicates ztailpacking */
- } else {
- pcl->index = erofs_blknr(sb, map->m_pa);
+ if (!pcl->from_meta) {
while (1) {
xa_lock(&sbi->managed_pslots);
- pre = __xa_cmpxchg(&sbi->managed_pslots, pcl->index,
+ pre = __xa_cmpxchg(&sbi->managed_pslots, pcl->pos,
NULL, pcl, GFP_KERNEL);
if (!pre || xa_is_err(pre) || z_erofs_get_pcluster(pre)) {
xa_unlock(&sbi->managed_pslots);
@@ -790,8 +791,7 @@ static int z_erofs_register_pcluster(struct z_erofs_decompress_frontend *fe)
goto err_out;
}
}
- fe->owned_head = &pcl->next;
- fe->pcl = pcl;
+ fe->head = fe->pcl = pcl;
return 0;
err_out:
@@ -800,24 +800,23 @@ err_out:
return err;
}
-static int z_erofs_pcluster_begin(struct z_erofs_decompress_frontend *fe)
+static int z_erofs_pcluster_begin(struct z_erofs_frontend *fe)
{
struct erofs_map_blocks *map = &fe->map;
struct super_block *sb = fe->inode->i_sb;
- erofs_blk_t blknr = erofs_blknr(sb, map->m_pa);
struct z_erofs_pcluster *pcl = NULL;
int ret;
DBG_BUGON(fe->pcl);
/* must be Z_EROFS_PCLUSTER_TAIL or pointed to previous pcluster */
- DBG_BUGON(fe->owned_head == Z_EROFS_PCLUSTER_NIL);
+ DBG_BUGON(!fe->head);
if (!(map->m_flags & EROFS_MAP_META)) {
while (1) {
rcu_read_lock();
- pcl = xa_load(&EROFS_SB(sb)->managed_pslots, blknr);
+ pcl = xa_load(&EROFS_SB(sb)->managed_pslots, map->m_pa);
if (!pcl || z_erofs_get_pcluster(pcl)) {
- DBG_BUGON(pcl && blknr != pcl->index);
+ DBG_BUGON(pcl && map->m_pa != pcl->pos);
rcu_read_unlock();
break;
}
@@ -838,10 +837,9 @@ static int z_erofs_pcluster_begin(struct z_erofs_decompress_frontend *fe)
if (ret == -EEXIST) {
mutex_lock(&fe->pcl->lock);
/* check if this pcluster hasn't been linked into any chain. */
- if (cmpxchg(&fe->pcl->next, Z_EROFS_PCLUSTER_NIL,
- fe->owned_head) == Z_EROFS_PCLUSTER_NIL) {
+ if (!cmpxchg(&fe->pcl->next, NULL, fe->head)) {
/* .. so it can be attached to our submission chain */
- fe->owned_head = &fe->pcl->next;
+ fe->head = fe->pcl;
fe->mode = Z_EROFS_PCLUSTER_FOLLOWED;
} else { /* otherwise, it belongs to an inflight chain */
fe->mode = Z_EROFS_PCLUSTER_INFLIGHT;
@@ -852,13 +850,13 @@ static int z_erofs_pcluster_begin(struct z_erofs_decompress_frontend *fe)
z_erofs_bvec_iter_begin(&fe->biter, &fe->pcl->bvset,
Z_EROFS_INLINE_BVECS, fe->pcl->vcnt);
- if (!z_erofs_is_inline_pcluster(fe->pcl)) {
+ if (!fe->pcl->from_meta) {
/* bind cache first when cached decompression is preferred */
z_erofs_bind_cache(fe);
} else {
void *mptr;
- mptr = erofs_read_metabuf(&map->buf, sb, map->m_pa, EROFS_NO_KMAP);
+ mptr = erofs_read_metabuf(&map->buf, sb, map->m_pa, false);
if (IS_ERR(mptr)) {
ret = PTR_ERR(mptr);
erofs_err(sb, "failed to get inline data %d", ret);
@@ -874,14 +872,9 @@ static int z_erofs_pcluster_begin(struct z_erofs_decompress_frontend *fe)
return 0;
}
-/*
- * keep in mind that no referenced pclusters will be freed
- * only after a RCU grace period.
- */
static void z_erofs_rcu_callback(struct rcu_head *head)
{
- z_erofs_free_pcluster(container_of(head,
- struct z_erofs_pcluster, rcu));
+ z_erofs_free_pcluster(container_of(head, struct z_erofs_pcluster, rcu));
}
static bool __erofs_try_to_release_pcluster(struct erofs_sb_info *sbi,
@@ -902,7 +895,7 @@ static bool __erofs_try_to_release_pcluster(struct erofs_sb_info *sbi,
* It's impossible to fail after the pcluster is freezed, but in order
* to avoid some race conditions, add a DBG_BUGON to observe this.
*/
- DBG_BUGON(__xa_erase(&sbi->managed_pslots, pcl->index) != pcl);
+ DBG_BUGON(__xa_erase(&sbi->managed_pslots, pcl->pos) != pcl);
lockref_mark_dead(&pcl->lockref);
return true;
@@ -923,8 +916,7 @@ static bool erofs_try_to_release_pcluster(struct erofs_sb_info *sbi,
return free;
}
-unsigned long z_erofs_shrink_scan(struct erofs_sb_info *sbi,
- unsigned long nr_shrink)
+unsigned long z_erofs_shrink_scan(struct erofs_sb_info *sbi, unsigned long nr)
{
struct z_erofs_pcluster *pcl;
unsigned long index, freed = 0;
@@ -937,7 +929,7 @@ unsigned long z_erofs_shrink_scan(struct erofs_sb_info *sbi,
xa_unlock(&sbi->managed_pslots);
++freed;
- if (!--nr_shrink)
+ if (!--nr)
return freed;
xa_lock(&sbi->managed_pslots);
}
@@ -966,7 +958,7 @@ static void z_erofs_put_pcluster(struct erofs_sb_info *sbi,
call_rcu(&pcl->rcu, z_erofs_rcu_callback);
}
-static void z_erofs_pcluster_end(struct z_erofs_decompress_frontend *fe)
+static void z_erofs_pcluster_end(struct z_erofs_frontend *fe)
{
struct z_erofs_pcluster *pcl = fe->pcl;
@@ -979,13 +971,9 @@ static void z_erofs_pcluster_end(struct z_erofs_decompress_frontend *fe)
if (fe->candidate_bvpage)
fe->candidate_bvpage = NULL;
- /*
- * if all pending pages are added, don't hold its reference
- * any longer if the pcluster isn't hosted by ourselves.
- */
+ /* Drop refcount if it doesn't belong to our processing chain */
if (fe->mode < Z_EROFS_PCLUSTER_FOLLOWED_NOINPLACE)
z_erofs_put_pcluster(EROFS_I_SB(fe->inode), pcl, false);
-
fe->pcl = NULL;
}
@@ -1003,7 +991,7 @@ static int z_erofs_read_fragment(struct super_block *sb, struct folio *folio,
buf.mapping = packed_inode->i_mapping;
for (; cur < end; cur += cnt, pos += cnt) {
cnt = min(end - cur, sb->s_blocksize - erofs_blkoff(sb, pos));
- src = erofs_bread(&buf, pos, EROFS_KMAP);
+ src = erofs_bread(&buf, pos, true);
if (IS_ERR(src)) {
erofs_put_metabuf(&buf);
return PTR_ERR(src);
@@ -1014,7 +1002,7 @@ static int z_erofs_read_fragment(struct super_block *sb, struct folio *folio,
return 0;
}
-static int z_erofs_scan_folio(struct z_erofs_decompress_frontend *f,
+static int z_erofs_scan_folio(struct z_erofs_frontend *f,
struct folio *folio, bool ra)
{
struct inode *const inode = f->inode;
@@ -1046,7 +1034,7 @@ static int z_erofs_scan_folio(struct z_erofs_decompress_frontend *f,
if (!(map->m_flags & EROFS_MAP_MAPPED)) {
folio_zero_segment(folio, cur, end);
tight = false;
- } else if (map->m_flags & EROFS_MAP_FRAGMENT) {
+ } else if (map->m_flags & __EROFS_MAP_FRAGMENT) {
erofs_off_t fpos = offset + cur - map->m_la;
err = z_erofs_read_fragment(inode->i_sb, folio, cur,
@@ -1086,8 +1074,6 @@ static int z_erofs_scan_folio(struct z_erofs_decompress_frontend *f,
break;
erofs_onlinefolio_split(folio);
- if (f->pcl->pageofs_out != (map->m_la & ~PAGE_MASK))
- f->pcl->multibases = true;
if (f->pcl->length < offset + end - map->m_la) {
f->pcl->length = offset + end - map->m_la;
f->pcl->pageofs_out = map->m_la & ~PAGE_MASK;
@@ -1105,7 +1091,7 @@ static int z_erofs_scan_folio(struct z_erofs_decompress_frontend *f,
tight = (bs == PAGE_SIZE);
}
} while ((end = cur) > 0);
- erofs_onlinefolio_end(folio, err);
+ erofs_onlinefolio_end(folio, err, false);
return err;
}
@@ -1129,11 +1115,10 @@ static bool z_erofs_page_is_invalidated(struct page *page)
return !page_folio(page)->mapping && !z_erofs_is_shortlived_page(page);
}
-struct z_erofs_decompress_backend {
+struct z_erofs_backend {
struct page *onstack_pages[Z_EROFS_ONSTACK_PAGES];
struct super_block *sb;
struct z_erofs_pcluster *pcl;
-
/* pages with the longest decompressed length for deduplication */
struct page **decompressed_pages;
/* pages to keep the compressed data */
@@ -1142,6 +1127,8 @@ struct z_erofs_decompress_backend {
struct list_head decompressed_secondary_bvecs;
struct page **pagepool;
unsigned int onstack_used, nr_pages;
+ /* indicate if temporary copies should be preserved for later use */
+ bool keepxcpy;
};
struct z_erofs_bvec_item {
@@ -1149,21 +1136,23 @@ struct z_erofs_bvec_item {
struct list_head list;
};
-static void z_erofs_do_decompressed_bvec(struct z_erofs_decompress_backend *be,
+static void z_erofs_do_decompressed_bvec(struct z_erofs_backend *be,
struct z_erofs_bvec *bvec)
{
+ int poff = bvec->offset + be->pcl->pageofs_out;
struct z_erofs_bvec_item *item;
- unsigned int pgnr;
-
- if (!((bvec->offset + be->pcl->pageofs_out) & ~PAGE_MASK) &&
- (bvec->end == PAGE_SIZE ||
- bvec->offset + bvec->end == be->pcl->length)) {
- pgnr = (bvec->offset + be->pcl->pageofs_out) >> PAGE_SHIFT;
- DBG_BUGON(pgnr >= be->nr_pages);
- if (!be->decompressed_pages[pgnr]) {
- be->decompressed_pages[pgnr] = bvec->page;
+ struct page **page;
+
+ if (!(poff & ~PAGE_MASK) && (bvec->end == PAGE_SIZE ||
+ bvec->offset + bvec->end == be->pcl->length)) {
+ DBG_BUGON((poff >> PAGE_SHIFT) >= be->nr_pages);
+ page = be->decompressed_pages + (poff >> PAGE_SHIFT);
+ if (!*page) {
+ *page = bvec->page;
return;
}
+ } else {
+ be->keepxcpy = true;
}
/* (cold path) one pcluster is requested multiple times */
@@ -1172,8 +1161,7 @@ static void z_erofs_do_decompressed_bvec(struct z_erofs_decompress_backend *be,
list_add(&item->list, &be->decompressed_secondary_bvecs);
}
-static void z_erofs_fill_other_copies(struct z_erofs_decompress_backend *be,
- int err)
+static void z_erofs_fill_other_copies(struct z_erofs_backend *be, int err)
{
unsigned int off0 = be->pcl->pageofs_out;
struct list_head *p, *n;
@@ -1208,13 +1196,13 @@ static void z_erofs_fill_other_copies(struct z_erofs_decompress_backend *be,
cur += len;
}
kunmap_local(dst);
- erofs_onlinefolio_end(page_folio(bvi->bvec.page), err);
+ erofs_onlinefolio_end(page_folio(bvi->bvec.page), err, true);
list_del(p);
kfree(bvi);
}
}
-static void z_erofs_parse_out_bvecs(struct z_erofs_decompress_backend *be)
+static void z_erofs_parse_out_bvecs(struct z_erofs_backend *be)
{
struct z_erofs_pcluster *pcl = be->pcl;
struct z_erofs_bvec_iter biter;
@@ -1239,8 +1227,7 @@ static void z_erofs_parse_out_bvecs(struct z_erofs_decompress_backend *be)
z_erofs_put_shortlivedpage(be->pagepool, old_bvpage);
}
-static int z_erofs_parse_in_bvecs(struct z_erofs_decompress_backend *be,
- bool *overlapped)
+static int z_erofs_parse_in_bvecs(struct z_erofs_backend *be, bool *overlapped)
{
struct z_erofs_pcluster *pcl = be->pcl;
unsigned int pclusterpages = z_erofs_pclusterpages(pcl);
@@ -1259,7 +1246,7 @@ static int z_erofs_parse_in_bvecs(struct z_erofs_decompress_backend *be,
}
be->compressed_pages[i] = page;
- if (z_erofs_is_inline_pcluster(pcl) ||
+ if (pcl->from_meta ||
erofs_folio_is_managed(EROFS_SB(be->sb), page_folio(page))) {
if (!PageUptodate(page))
err = -EIO;
@@ -1275,8 +1262,7 @@ static int z_erofs_parse_in_bvecs(struct z_erofs_decompress_backend *be,
return err;
}
-static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
- int err)
+static int z_erofs_decompress_pcluster(struct z_erofs_backend *be, int err)
{
struct erofs_sb_info *const sbi = EROFS_SB(be->sb);
struct z_erofs_pcluster *pcl = be->pcl;
@@ -1323,6 +1309,8 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
.sb = be->sb,
.in = be->compressed_pages,
.out = be->decompressed_pages,
+ .inpages = pclusterpages,
+ .outpages = be->nr_pages,
.pageofs_in = pcl->pageofs_in,
.pageofs_out = pcl->pageofs_out,
.inputsize = pcl->pclustersize,
@@ -1330,13 +1318,13 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
.alg = pcl->algorithmformat,
.inplace_io = overlapped,
.partial_decoding = pcl->partial,
- .fillgaps = pcl->multibases,
+ .fillgaps = be->keepxcpy,
.gfp = pcl->besteffort ? GFP_KERNEL :
GFP_NOWAIT | __GFP_NORETRY
}, be->pagepool);
/* must handle all compressed pages before actual file pages */
- if (z_erofs_is_inline_pcluster(pcl)) {
+ if (pcl->from_meta) {
page = pcl->compressed_bvecs[0].page;
WRITE_ONCE(pcl->compressed_bvecs[0].page, NULL);
put_page(page);
@@ -1367,7 +1355,7 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
DBG_BUGON(z_erofs_page_is_invalidated(page));
if (!z_erofs_is_shortlived_page(page)) {
- erofs_onlinefolio_end(page_folio(page), err);
+ erofs_onlinefolio_end(page_folio(page), err, true);
continue;
}
if (pcl->algorithmformat != Z_EROFS_COMPRESSION_LZ4) {
@@ -1387,16 +1375,15 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
pcl->length = 0;
pcl->partial = true;
- pcl->multibases = false;
pcl->besteffort = false;
pcl->bvset.nextpage = NULL;
pcl->vcnt = 0;
/* pcluster lock MUST be taken before the following line */
- WRITE_ONCE(pcl->next, Z_EROFS_PCLUSTER_NIL);
+ WRITE_ONCE(pcl->next, NULL);
mutex_unlock(&pcl->lock);
- if (z_erofs_is_inline_pcluster(pcl))
+ if (pcl->from_meta)
z_erofs_free_pcluster(pcl);
else
z_erofs_put_pcluster(sbi, pcl, try_free);
@@ -1406,21 +1393,19 @@ static int z_erofs_decompress_pcluster(struct z_erofs_decompress_backend *be,
static int z_erofs_decompress_queue(const struct z_erofs_decompressqueue *io,
struct page **pagepool)
{
- struct z_erofs_decompress_backend be = {
+ struct z_erofs_backend be = {
.sb = io->sb,
.pagepool = pagepool,
.decompressed_secondary_bvecs =
LIST_HEAD_INIT(be.decompressed_secondary_bvecs),
+ .pcl = io->head,
};
- z_erofs_next_pcluster_t owned = io->head;
+ struct z_erofs_pcluster *next;
int err = io->eio ? -EIO : 0;
- while (owned != Z_EROFS_PCLUSTER_TAIL) {
- DBG_BUGON(owned == Z_EROFS_PCLUSTER_NIL);
-
- be.pcl = container_of(owned, struct z_erofs_pcluster, next);
- owned = READ_ONCE(be.pcl->next);
-
+ for (; be.pcl != Z_EROFS_PCLUSTER_TAIL; be.pcl = next) {
+ DBG_BUGON(!be.pcl);
+ next = READ_ONCE(be.pcl->next);
err = z_erofs_decompress_pcluster(&be, err) ?: err;
}
return err;
@@ -1486,7 +1471,7 @@ static void z_erofs_decompress_kickoff(struct z_erofs_decompressqueue *io,
}
static void z_erofs_fill_bio_vec(struct bio_vec *bvec,
- struct z_erofs_decompress_frontend *f,
+ struct z_erofs_frontend *f,
struct z_erofs_pcluster *pcl,
unsigned int nr,
struct address_space *mc)
@@ -1513,12 +1498,8 @@ repeat:
DBG_BUGON(z_erofs_is_shortlived_page(bvec->bv_page));
folio = page_folio(zbv.page);
- /*
- * Handle preallocated cached folios. We tried to allocate such folios
- * without triggering direct reclaim. If allocation failed, inplace
- * file-backed folios will be used instead.
- */
- if (folio->private == (void *)Z_EROFS_PREALLOCATED_PAGE) {
+ /* For preallocated managed folios, add them to page cache here */
+ if (folio->private == Z_EROFS_PREALLOCATED_FOLIO) {
tocache = true;
goto out_tocache;
}
@@ -1583,7 +1564,7 @@ out_allocfolio:
folio = page_folio(page);
out_tocache:
if (!tocache || bs != PAGE_SIZE ||
- filemap_add_folio(mc, folio, pcl->index + nr, gfp)) {
+ filemap_add_folio(mc, folio, (pcl->pos >> PAGE_SHIFT) + nr, gfp)) {
/* turn into a temporary shortlived folio (1 ref) */
folio->private = (void *)Z_EROFS_SHORTLIVED_PAGE;
return;
@@ -1630,18 +1611,13 @@ enum {
NR_JOBQUEUES,
};
-static void move_to_bypass_jobqueue(struct z_erofs_pcluster *pcl,
- z_erofs_next_pcluster_t qtail[],
- z_erofs_next_pcluster_t owned_head)
+static void z_erofs_move_to_bypass_queue(struct z_erofs_pcluster *pcl,
+ struct z_erofs_pcluster *next,
+ struct z_erofs_pcluster **qtail[])
{
- z_erofs_next_pcluster_t *const submit_qtail = qtail[JQ_SUBMIT];
- z_erofs_next_pcluster_t *const bypass_qtail = qtail[JQ_BYPASS];
-
WRITE_ONCE(pcl->next, Z_EROFS_PCLUSTER_TAIL);
-
- WRITE_ONCE(*submit_qtail, owned_head);
- WRITE_ONCE(*bypass_qtail, &pcl->next);
-
+ WRITE_ONCE(*qtail[JQ_SUBMIT], next);
+ WRITE_ONCE(*qtail[JQ_BYPASS], pcl);
qtail[JQ_BYPASS] = &pcl->next;
}
@@ -1670,15 +1646,15 @@ static void z_erofs_endio(struct bio *bio)
bio_put(bio);
}
-static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
+static void z_erofs_submit_queue(struct z_erofs_frontend *f,
struct z_erofs_decompressqueue *fgq,
bool *force_fg, bool readahead)
{
struct super_block *sb = f->inode->i_sb;
struct address_space *mc = MNGD_MAPPING(EROFS_SB(sb));
- z_erofs_next_pcluster_t qtail[NR_JOBQUEUES];
+ struct z_erofs_pcluster **qtail[NR_JOBQUEUES];
struct z_erofs_decompressqueue *q[NR_JOBQUEUES];
- z_erofs_next_pcluster_t owned_head = f->owned_head;
+ struct z_erofs_pcluster *pcl, *next;
/* bio is NULL initially, so no need to initialize last_{index,bdev} */
erofs_off_t last_pa;
unsigned int nr_bios = 0;
@@ -1694,33 +1670,31 @@ static void z_erofs_submit_queue(struct z_erofs_decompress_frontend *f,
qtail[JQ_SUBMIT] = &q[JQ_SUBMIT]->head;
/* by default, all need io submission */
- q[JQ_SUBMIT]->head = owned_head;
+ q[JQ_SUBMIT]->head = next = f->head;
do {
struct erofs_map_dev mdev;
- struct z_erofs_pcluster *pcl;
erofs_off_t cur, end;
struct bio_vec bvec;
unsigned int i = 0;
bool bypass = true;
- DBG_BUGON(owned_head == Z_EROFS_PCLUSTER_NIL);
- pcl = container_of(owned_head, struct z_erofs_pcluster, next);
- owned_head = READ_ONCE(pcl->next);
-
- if (z_erofs_is_inline_pcluster(pcl)) {
- move_to_bypass_jobqueue(pcl, qtail, owned_head);
+ pcl = next;
+ next = READ_ONCE(pcl->next);
+ if (pcl->from_meta) {
+ z_erofs_move_to_bypass_queue(pcl, next, qtail);
continue;
}
/* no device id here, thus it will always succeed */
mdev = (struct erofs_map_dev) {
- .m_pa = erofs_pos(sb, pcl->index),
+ .m_pa = round_down(pcl->pos, sb->s_blocksize),
};
(void)erofs_map_dev(sb, &mdev);
cur = mdev.m_pa;
- end = cur + pcl->pclustersize;
+ end = round_up(cur + pcl->pageofs_in + pcl->pclustersize,
+ sb->s_blocksize);
do {
bvec.bv_page = NULL;
if (bio && (cur != last_pa ||
@@ -1764,7 +1738,8 @@ drain_io:
bio = bio_alloc(mdev.m_bdev, BIO_MAX_VECS,
REQ_OP_READ, GFP_NOIO);
bio->bi_end_io = z_erofs_endio;
- bio->bi_iter.bi_sector = cur >> 9;
+ bio->bi_iter.bi_sector =
+ (mdev.m_dif->fsoff + cur) >> 9;
bio->bi_private = q[JQ_SUBMIT];
if (readahead)
bio->bi_opf |= REQ_RAHEAD;
@@ -1781,8 +1756,8 @@ drain_io:
if (!bypass)
qtail[JQ_SUBMIT] = &pcl->next;
else
- move_to_bypass_jobqueue(pcl, qtail, owned_head);
- } while (owned_head != Z_EROFS_PCLUSTER_TAIL);
+ z_erofs_move_to_bypass_queue(pcl, next, qtail);
+ } while (next != Z_EROFS_PCLUSTER_TAIL);
if (bio) {
if (erofs_is_fileio_mode(EROFS_SB(sb)))
@@ -1806,17 +1781,16 @@ drain_io:
z_erofs_decompress_kickoff(q[JQ_SUBMIT], nr_bios);
}
-static int z_erofs_runqueue(struct z_erofs_decompress_frontend *f,
- unsigned int ra_folios)
+static int z_erofs_runqueue(struct z_erofs_frontend *f, unsigned int rapages)
{
struct z_erofs_decompressqueue io[NR_JOBQUEUES];
struct erofs_sb_info *sbi = EROFS_I_SB(f->inode);
- bool force_fg = z_erofs_is_sync_decompress(sbi, ra_folios);
+ bool force_fg = z_erofs_is_sync_decompress(sbi, rapages);
int err;
- if (f->owned_head == Z_EROFS_PCLUSTER_TAIL)
+ if (f->head == Z_EROFS_PCLUSTER_TAIL)
return 0;
- z_erofs_submit_queue(f, io, &force_fg, !!ra_folios);
+ z_erofs_submit_queue(f, io, &force_fg, !!rapages);
/* handle bypass queue (no i/o pclusters) immediately */
err = z_erofs_decompress_queue(&io[JQ_BYPASS], &f->pagepool);
@@ -1834,7 +1808,7 @@ static int z_erofs_runqueue(struct z_erofs_decompress_frontend *f,
* Since partial uptodate is still unimplemented for now, we have to use
* approximate readmore strategies as a start.
*/
-static void z_erofs_pcluster_readmore(struct z_erofs_decompress_frontend *f,
+static void z_erofs_pcluster_readmore(struct z_erofs_frontend *f,
struct readahead_control *rac, bool backmost)
{
struct inode *inode = f->inode;
@@ -1889,12 +1863,10 @@ static void z_erofs_pcluster_readmore(struct z_erofs_decompress_frontend *f,
static int z_erofs_read_folio(struct file *file, struct folio *folio)
{
struct inode *const inode = folio->mapping->host;
- struct z_erofs_decompress_frontend f = DECOMPRESS_FRONTEND_INIT(inode);
+ Z_EROFS_DEFINE_FRONTEND(f, inode, folio_pos(folio));
int err;
trace_erofs_read_folio(folio, false);
- f.headoffset = (erofs_off_t)folio->index << PAGE_SHIFT;
-
z_erofs_pcluster_readmore(&f, NULL, true);
err = z_erofs_scan_folio(&f, folio, false);
z_erofs_pcluster_readmore(&f, NULL, false);
@@ -1914,17 +1886,13 @@ static int z_erofs_read_folio(struct file *file, struct folio *folio)
static void z_erofs_readahead(struct readahead_control *rac)
{
struct inode *const inode = rac->mapping->host;
- struct z_erofs_decompress_frontend f = DECOMPRESS_FRONTEND_INIT(inode);
+ Z_EROFS_DEFINE_FRONTEND(f, inode, readahead_pos(rac));
+ unsigned int nrpages = readahead_count(rac);
struct folio *head = NULL, *folio;
- unsigned int nr_folios;
int err;
- f.headoffset = readahead_pos(rac);
-
+ trace_erofs_readahead(inode, readahead_index(rac), nrpages, false);
z_erofs_pcluster_readmore(&f, rac, true);
- nr_folios = readahead_count(rac);
- trace_erofs_readpages(inode, readahead_index(rac), nr_folios, false);
-
while ((folio = readahead_folio(rac))) {
folio->private = head;
head = folio;
@@ -1943,7 +1911,7 @@ static void z_erofs_readahead(struct readahead_control *rac)
z_erofs_pcluster_readmore(&f, rac, false);
z_erofs_pcluster_end(&f);
- (void)z_erofs_runqueue(&f, nr_folios);
+ (void)z_erofs_runqueue(&f, nrpages);
erofs_put_metabuf(&f.map.buf);
erofs_release_pages(&f.pagepool);
}
generated by cgit v1.2.3 (git 2.39.0) at 2025-08-21 08:36:17 +0300